Joint Grouping and Offloading in NOMA-Assisted Multi-MEC IoVT Systems

As a new development of Internet of Things (IoT), Internet of Video Things (IoVT) emerges to provide novel services based on video sensing, transmission, storage and analysis. However, IoVT also imposes massive computation and transmission on mobile edge computing (MEC) systems with a large amount of offloaded video data. To address this issue, in this paper, we propose a non-orthogonal multiple access (NOMA) assisted multi-MEC IoVT system. Although NOMA-assisted MEC systems have been proposed in existing studies, non-negligible inter-device interference caused by NOMA transmission has not been investigated, which is much more serious in multi-MEC IoVT systems. To combat them, we perform joint optimization on grouping and offloading in the proposed NOMA-assisted multi-MEC IoVT system. To achieve optimal performance, a utility minimization problem is formulated where the definition of utility leverages critical performance metrics including energy consumption and delay. We prove that this problem can be modeled as an exact potential game. Then, a selection algorithm is proposed to find the optimal strategies for each IoVT device by obtaining the Nash equilibrium. Specifically, power allocation for IoVT devices can be handled by a computation resource minimization problem. Simulation results demonstrate that the proposed algorithm achieves significant performance gains compared with existing schemes, i.e., at least 34.8% reduction in total utility and 14.5% less power consumption.